GenSight.AI has now run AI visibility benchmarks across two industries - Finance and Content Marketing - in both enterprise and influencer mode. Four benchmark runs. Sixty audited entities. Two completely different types of subject: organisations and people.
The data tells a consistent story across both industries, and it has significant implications for how GEO programmes should be designed, resourced, and measured depending on who the entity being optimised actually is.
The finding is this: AI knowledge systems treat a company and a person as fundamentally different types of entity. The signals that determine visibility are different, the gaps that limit citation are different, and the remediation required to close those gaps is different in kind, not just degree. Running the same GEO programme for a personal brand that you would run for an enterprise is not just suboptimal - it is solving the wrong problem with the wrong tools.
The Numbers Across Four Benchmarks
The gap between enterprise and influencer performance is consistent across both industries. In Finance, enterprise institutions averaged 68 against influencers averaging 56 - a 12-point gap. In Content Marketing, enterprise platforms averaged 66 against practitioners averaging 55 - an 11-point gap. Two different industries, two different groups of audited entities, the same structural divergence.
What makes this finding significant is not the size of the gap. It is the consistency. An 11 to 12 point enterprise advantage appearing across industries as structurally different as financial services and content marketing suggests this is not a sector-specific phenomenon. It is a systemic feature of how AI knowledge systems are built - and how differently they process organisational versus personal entities.
The metric-level data makes the structural difference even clearer. In both industries, enterprise entities lead on Citation Worthiness by 11 to 13 points. They lead on Retrieval Optimisation by 13 points. They lead on the score distribution that matters most: in Finance, seven of fifteen enterprise entities reached the high tier while zero influencers did. In Content Marketing, two enterprise platforms reached the high tier against one influencer.
Why the Knowledge Graph Treats Organisations and People Differently
The explanation for this gap begins with how AI knowledge systems were trained and how they structure their understanding of the world. The schema.org vocabulary - the structured data framework that underpins how machines declare and understand entities - has distinct types for Organisation and Person. These are not stylistic variations. They carry different properties, different verification signals, and different citation architectures.
An Organisation entity is verified through its products, its registered presence, its Wikidata entry, its review aggregator profiles, its structured data Schema, its third-party analyst coverage, and its comparison platform presence. These signals are largely digital-native and can be deployed systematically by a team.
A Person entity is verified through a different set of signals: Person Schema on an owned domain, the SameAs array linking verified social profiles, audio and video transcript availability, FAQ schema for authored concepts, entity graph presence through co-citation with established organisations, and the consistency of identity signals across multiple platforms. Most of these signals are individually lightweight. Together, they are the architecture that determines whether AI treats someone as a distinct, citable authority or as an ambiguous reference in someone else's content.
The practical implication is that the GEO checklist for a personal brand is not a shortened version of the enterprise checklist. It is a different checklist entirely - built around a different schema type, different verification pathways, and different content formats that earn citation in the personal brand context.
The Format Problem Is Personal Brand Specific
In the enterprise benchmarks, the gaps are largely metadata gaps. The content exists in machine-readable formats - crawlable web pages, structured HTML, text-based knowledge bases. What most enterprise entities are missing is the declaration layer: Organisation Schema, llms.txt directives, structured social proof, entity-mapped breadcrumbs. The content is there. The signals that tell AI how to interpret and attribute it are largely absent.
In the influencer benchmarks, the gap goes deeper. Significant portions of intellectual property do not exist in AI-readable formats at all. In Finance, the Market Analysis and Strategy sub-group averaged a Retrieval Optimisation score of 40 - the lowest of any sub-group across any benchmark. These are practitioners whose work lives in fund research reports, video analysis, podcast commentary, and long-form written essays. All of these formats require additional conversion work before AI can read them efficiently.
Across both influencer benchmarks - Finance and Content Marketing - not one of the thirty audited practitioners uses structured transcripts for their audio and video content. This is the single most universal gap in the entire dataset. Thirty individuals with significant intellectual property in spoken and recorded formats. Not one has converted that IP into the structured, machine-readable text format that AI retrieval systems require.
For an enterprise entity, this gap largely does not exist - companies publish predominantly in text-based web formats by default. For a personal brand whose primary medium is the podcast, the keynote, the YouTube video, or the recorded interview, the gap between what has been said and what AI can read is enormous.
Where the Sub-Group Data Confirms the Structural Split
The sub-group patterns across both industries reinforce the same structural conclusion. In the enterprise benchmarks, the highest-performing sub-groups are consistently the ones whose products require structured data architecture by design. In Finance, Digital-First FinTech averages 73 - companies built for digital-native customers produce structured, direct-answer content as a byproduct of their product design. In Content Marketing, Content Management platforms averaged 69 for the same reason.
In the influencer benchmarks, the highest-performing sub-groups are the ones whose content is most structured by nature or most grounded in owned web infrastructure. In Finance, Personal Finance and Investing Education averaged 60 - practitioners whose output includes books, structured website content, and direct-answer advice columns. In Content Marketing, the Content Creation and Business Growth group averaged 58 - practitioners with clear product offerings and structured owned web presences.
The lowest-performing influencer sub-groups in both industries are those whose intellectual property is most analytical, most nuanced, and most likely to live in long-form formats that resist machine extraction. Finance Market Analysis and Strategy averaged 51. Content Marketing Strategy and Thought Leadership averaged 51. The same number, in different industries, for structurally similar reasons.
The pattern is not random. It is a consistent signal that the type of content a practitioner produces, and the format it naturally lives in, are the primary determinants of AI visibility - independent of the quality of the ideas or the size of the audience.
The Goldman Sachs and Seth Godin Problem
The most instructive individual data points across the four benchmarks are Goldman Sachs in Finance enterprise (47) and Seth Godin in Content Marketing influencer (the paradox score discussed in the Content Marketing benchmark). Both represent the same phenomenon at different ends of the entity spectrum.
Goldman Sachs is arguably the most prestigious financial institution in the world by human reputation metrics. It scored 47 because the majority of its content is published in formats - gated PDFs, dense regulatory documents, press releases - that are poorly structured for AI extraction. The authority is real. The infrastructure to express that authority in machine-readable form does not exist in the current digital estate.
Seth Godin leads the Content Marketing Thought Leadership sub-group despite having weaker structured signals than his overall fame would suggest. He scores higher than his peers in that sub-group partly because his daily blog provides an unusually large volume of text for AI to index, even if that text is not structured for optimal retrieval. But his Retrieval Optimisation score is among the lowest in the influencer benchmark because the format of that content - short, aphoristic, non-structured blog posts - does not give AI systems the scaffolding they need to extract and attribute specific claims efficiently.
Both cases illustrate the same principle: human authority and AI citability are not the same currency, and they cannot be exchanged at par. Converting one into the other requires deliberate infrastructure work that neither Goldman Sachs nor most personal brand practitioners have systematically undertaken.
Why the Remediation Is Different in Kind
The data across these four benchmarks points to a consistent conclusion: the interventions required to close the personal brand AI visibility gap do not appear on any enterprise GEO roadmap, and cannot be derived from enterprise GEO logic.
Enterprise GEO remediation operates primarily at the declaration layer - adding structured signals to content that already exists in the right format. The work is largely technical and can be systematically prioritised by a team against a defined checklist of organisation-level signals.
Personal brand GEO remediation operates across fundamentally different dimensions. The schema type is different. The content format requirements are different. The verification pathway - how AI establishes that a person is a distinct, authoritative entity rather than an ambiguous reference - is built from signals that have no enterprise equivalent. And the intellectual property problem that most personal brand practitioners face - significant portions of their best work existing in formats AI systems cannot read - requires a category of solution that enterprise GEO does not address at all.
The result is that the gap between enterprise and personal brand AI visibility is not primarily a resource gap or an effort gap. It is a programme design gap. An enterprise GEO programme applied to a personal brand will improve some signals and miss the ones that matter most for Person-type entity verification. The checklist is not shorter. It is different.
The Consistent 12-Point Gap and What It Predicts
The 11 to 12 point enterprise advantage across both industries will compound over time if personal brand GEO is not treated as a distinct discipline. As AI becomes a more significant interface for professional discovery - for buyers researching expertise, for audiences finding voices to follow, for organisations evaluating thought leaders - the default answers AI generates will increasingly reflect the structural advantages that enterprise entities hold.
Enterprise organisations have teams that can implement structured data, content operations that produce machine-readable output by default, and technical roadmaps that can prioritise GEO signals systematically. Individual practitioners have ideas, audiences, and limited bandwidth. The capacity gap compounds the infrastructure gap.
But the infrastructure gap itself is the more fundamental issue - and it is the one that a discipline-specific personal brand GEO programme directly addresses. The thirty influencers across these two benchmarks have the underlying authority to score significantly higher than they currently do. Topical Authority averages 81 in Finance and 83 in Content Marketing. These are people AI systems recognise as experts. The gap between that recognition and reliable citation is not an authority gap. It is an infrastructure gap. And unlike authority gaps, infrastructure gaps have deterministic solutions.
The data across four benchmarks, two industries, and sixty entities points to the same conclusion: the GEO programme that serves a brand does not serve a person. Building the discipline to address both, separately and correctly, is what the next phase of AI visibility strategy requires.
Data derived from the GenSight.AI Industry Benchmark Index by running deterministic vector gap analyses across the top entities. Bulk indexing capabilities will be available to partners on the Agency tier.